Research on Monitoring Topping Time of Cotton Based on AdaBoost+Decision Tree

Table 4

The regression of the combinations band and the cotton of plant height, buds, and fruiting branches with decision tree, random forests, and AdaBoost + decision tree.

Input band	Fitting index	Method	Fitting result
			R^2	RMSEP
730 + 790 nm	Cotton plant height	Decision tree	0.90	0.91
		Random forests	0.90	0.89
		AdaBoost + decision tree	0.90	0.86
	Number of buds	Decision tree	0.89	0.30
		Random forests	0.88	0.94
		AdaBoost + decision tree	0.91	0.95
	Number of fruiting branches	Decision tree	0.90	0.13
		Random forests	0.90	0.79
		AdaBoost + decision tree	0.95	0.99
550 + 730+790 nm	Cotton plant height	Decision tree	0.91	0.91
		Random forests	0.88	0.51
		AdaBoost + decision tree	0.96	0.40
	Number of flower buds	Decision tree	0.91	0.98
		Random forests	0.81	0.68
		AdaBoost + decision tree	0.84	0.49
	Number of fruiting branches	Decision tree	0.91	0.75
		Random forests	0.86	0.93
		AdaBoost + decision tree	0.97	0.54
Full band spectra	Cotton plant height	Decision tree	0.89	0.87
		Random forests	0.89	0.88
		AdaBoost + decision tree	0.93	0.23
	Number of buds	Decision tree	0.89	0.38
		Random forests	0.88	0.92
		AdaBoost + decision tree	0.90	0.97
	Number of fruiting branches	Decision tree	0.89	0.96
		Random forests	0.90	0.97
		AdaBoost + decision tree	0.95	0.52